Forem: Fatih Baltaci

AI Hallucinations

Fatih Baltaci — Wed, 04 Mar 2026 16:19:29 +0000

AI hallucinations aren't just annoying, they're a legal and business liability.

A lawyer was fined $5,000 for submitting six fabricated case citations generated by ChatGPT.

Air Canada was ordered to pay a customer after its chatbot invented a discount policy that never existed.

Google's AI Overviews told users to put glue on pizza.

And employees spend 4.5 hours per week just correcting AI mistakes.

These aren't edge cases anymore. There are 979 documented hallucination cases across 31 countries.

The EU AI Act is rolling out enforcement. 42 U.S. state attorneys general have warned major AI companies to fix this or face legal action.

"The AI said so" is no longer an excuse, it's becoming a liability.

So how do you actually prevent hallucinations?

At Gurubase, we built a multi-layer defense system:

Source grounding: answers come only from your verified knowledge base
Multi-query vector search to catch relevant context despite wording differences
LLM-based evaluation scoring context relevance on a 0-1 scale
Trust scores visible to users (green, yellow, red)
Source attribution with direct links to original documents
And when the system isn't confident? It says "I don't know" instead of making things up

The full breakdown with real-world examples, regulatory details, and technical architecture is in our latest blog post.

AI Hallucinations - Gurubase Blog

AI hallucinations have led to court fines, wrong diagnoses, and costly business errors. Learn why they happen, see real cases, and how to prevent them.

gurubase.io

Gurubase - AI-Powered Q&A Assistants for Any Topic

Fatih Baltaci — Tue, 21 Jan 2025 22:20:59 +0000

Gurubase: An open-source (Apache-2.0) RAG system to create your Personal Q&A assistants 🌟

💡 Features:
🤖 AI-Powered Q&A: Advanced LLM-based answering with instant evaluation to minimize hallucinations.
🔄 RAG System: Retrieval-Augmented Generation for accurate, context-aware responses.
📚 Multiple Data Sources: Index websites, PDFs, YouTube videos, and GitHub repositories.
🔌 Easy Integration: Embeddable widget for websites. Discord and Slack Bots are coming soon.
🎯 Custom Gurus: Build specialized AI assistants for specific topics.
🔄 Real-Time Updates: Keep data fresh by reindexing with a single click.
⛬ Binge Mode: Visualize and personalize your learning path while interacting with your Guru.
🛠 Self-Hosted Option: Full control over your deployment

Create Your Personal AI Guru – Add Data Sources (Websites, YouTube, PDFs, GitHub) and Get Instant Answers!

GitHub Repo: https://github.com/Gurubase/gurubase

Unveiling eBPF Verifier Errors - Episode 1

Fatih Baltaci — Tue, 16 Jan 2024 14:40:04 +0000

Liquid error: internal

A Comprehensive Guide to Using Postman Collections for Effective API Endpoint Load Testing

Fatih Baltaci — Tue, 16 Jan 2024 14:36:46 +0000

Liquid error: internal

Effortless Kubernetes Monitoring and Bottleneck Detection using eBPF 🐝

Fatih Baltaci — Fri, 03 Nov 2023 14:00:43 +0000

Liquid error: internal

Ddosify - Canva of Observability

Fatih Baltaci — Thu, 05 Oct 2023 12:30:31 +0000

Ddosify is a magic wand that instantly spots glitches and guarantees the smooth performance of your infrastructure and application while saving you time and money. Ddosify Platform includes Performance Testing and Kubernetes Observability capabilities. It uniquely integrates these two parts and effortlessly spots the performance issues.

We are aware that there are lots of open-source/commercial tools out there. However, we believe that monitoring the clusters and extracting actionable insights requires deep know-how about the tools/domain. We mainly focused on this problem.

🐝 Our eBPF agent (Alaz) is installed on your K8S cluster as DaemonSet. Thanks to eBPF - our agent collects traces directly from Linux kernels. This means there's no need for sidecars, instrumentations, or service restarts.

🎯 The UI not only visualizes data but also provides actionable insights. Using the Service Map, you can:

View latencies and RPS between services.
Detect zombie services and underperforming SQL queries.
Monitor golden signals, such as 5xx status codes.

In addition, Alaz (our eBPF agent) can capture system resources like CPU, Memory, Disk, and Network through the Prometheus Node Exporter, which is embedded in the agent.

🚀 Setting up is straightforward: just install Alaz as a DaemonSet, and the platform will handle the rest.

Finally, the combination of Alaz and Ddosify Performance Testing makes it possible to do load testing and simultaneously monitor the system to find bottlenecks instantly.

Your feedback would be greatly appreciated! 🙏

getanteon / anteon

Anteon (formerly Ddosify): eBPF-based Kubernetes Monitoring and Performance Testing

eBPF-powered Kubernetes Monitoring and Performance Testing

Anteon automatically generates Service Map of your K8s cluster without code instrumentation or sidecars. So you can easily find the bottlenecks in your system. Red lines indicate the high latency between services.

Live Demo • Documentation • Discord

🐝 What is Anteon?

Anteon (formerly Ddosify) is an open-source, eBPF-based Kubernetes Monitoring and Performance Testing platform.

🔎 Kubernetes Monitoring

Automatic Service Map Creation: Anteon automatically creates a service map of your cluster without code instrumentation or sidecars. So you can easily find the bottlenecks in your system.
Performance Insights: It helps you spot issues like services taking too long to respond or slow SQL queries.
Real-Time Metrics: The platform tracks and displays live data on your cluster instances CPU, memory, disk, and network usage.
Ease of Use: You don't need to change any code, restart services, or add extra components (like sidecars) to get…

View on GitHub

Performance Testing - What? Why? How? When? (with Ddosify Examples)

Fatih Baltaci — Mon, 29 May 2023 16:16:29 +0000

Performance Testing - What? Why? How? When? (with Ddosify Examples)

More: https://github.com/ddosify/ddosify

I built a ChatGPT Clone with React and Django, and GPT-4 model wrote most of the code! 🤖

Fatih Baltaci — Fri, 24 Mar 2023 15:12:18 +0000

Twitter Thread:
// Detect dark theme var iframe = document.getElementById('tweet-1639278279197614083-524'); if (document.body.className.includes('dark-theme')) { iframe.src = "https://platform.twitter.com/embed/Tweet.html?id=1639278279197614083&theme=dark" }

fatihbaltaci / chatgpt-clone

Self Hosted ChatGPT Clone: A Full-Stack Conversational AI Platform powered by React, Django REST Framework, and PostgreSQL and OpenAI's GPT-3.5 Model. Most of the code is written by GPT-4 model.

🤖 Self Hosted ChatGPT Clone

ChatGPT Clone: Self Hosted, Full-Stack Conversational AI Platform powered by React, Django REST Framework, and PostgreSQL and OpenAI's GPT-3.5 API. Most of the code is written by GPT-4 model.

📖 Table of Contents

🚀 Features

Self Hosted: You can host the project on your own server or local machine. All the data and OpenAI API Key are stored on your server.
Full-Stack: The project is a full-stack project that includes a React frontend, Django REST Framework backend, and PostgreSQL database.
Conversational AI: The project uses OpenAI's GPT-3.5 API to generate responses to user inputs.
Docker Compose: The project uses Docker Compose to run the project in containers.
Django Admin Panel: The project includes a Django Admin Panel to manage the project.
Supported OS: The project…

View on GitHub

Testing the Performance of User Authentication Flow

Fatih Baltaci — Sat, 14 Jan 2023 13:41:36 +0000

Hi Everyone,

I just want to share a blog article that explains how to test the performance of the user authentication flow with Ddosify open source load engine.

Liquid error: internal

Ddosify Latency Testing GitHub Action

Fatih Baltaci — Mon, 09 Jan 2023 10:17:44 +0000

In this article, we will demonstrate how to use the Ddosify GitHub Action to periodically test the latency of our target endpoints from 60+ cities worldwide. By configuring the action to run on a schedule, we can ensure that the performance of our locations is consistently monitored. If the latency of any of our locations exceeds the expected value, we will receive an email notification alerting us to the issue. By using this action, we can proactively identify and address potential performance issues before they impact our users.

📌 Check out the blog post:

Ddosify Latency Testing GitHub Action | Ddosify Blog

Ddosify Latency Testing GitHub Action

ddosify.com

📌 Check out the Ddosify Latency Action:

ddosify / ddosify-latency-action

Ddosify Latency Testing Action

Ddosify's Latency Observation Action allows you to measure the latency of endpoints from 60+ cities around the world With this action, you can track the performance of your endpoints over time and ensure that they are meeting your desired latency targets. Whether you are running a web application, a mobile backend, or any other type of service, this action can help you monitor the latency for the best user experience.

It is easy to use - just add it to your workflow and configure the request details. You can specify the Target URL, Locations and Fail If scenarios to fail the pipeline based on latencies. This action uses Ddosify Latency Testing API under the hood.

⚠️ This action uses Ddosify Cloud API Key that you can store in Github Secrets.

Inputs

input	required	default	example	description
api_key	true	Ddosify Cloud API Key. Instructions are

…

View on GitHub

Latency testing made easy with Ddosify Cloud

Fatih Baltaci — Mon, 12 Dec 2022 14:22:22 +0000

⌛ Latency is crucial, from low-ping gaming to a smooth fintech experience. Periodically testing endpoints' latency across the globe is time-consuming and costly.

Ddosify Cloud has a brand new feature that is capable of city-level latency testing from 60+ cities worldwide.
✅ Select cities and create a periodic latency test.
✅ By Latency API, create a decision-making system based on latency data

What about "If latency is higher than 50ms on Tokyo, ping me on Slack/Email/PagerDuty"? This will also be available soon.

Let’s try: https://app.ddosify.com/

Ddosify Open-source: https://github.com/ddosify/ddosify

Django Rest Framework Throttling with API Key

Fatih Baltaci — Tue, 23 Aug 2022 15:46:07 +0000

My new blog post about Django Rest Framework throttling is ready 🚀 In this blog, we test the throttling with different Django cache configurations (LocMemCache and Redis) and deploy it to Kubernetes. We used the open-source Ddosify load generator for the load test and Docker Desktop for single-node Kubernetes.

https://ddosify.com/blog/django-restframework-throttling-with-apikey